CN106226765A - A kind of building layout formation method and system - Google Patents

Abstract

The open a kind of building layout formation method of the present invention and system, first with the geometry site of body of wall Yu antenna, estimate body of wall parameter, calculate electromagnetic wave and penetrate body of wall time delay, the time delay truly penetrating body of wall is used as follow-up imaging compensating.Body of wall quantity in whole building interior space is few, has openness, utilizes body of wall to have bulk, successional feature simultaneously.Use distributed compression sensing method, to division subspace, target range territory, by calculated true propagation time delay in subspace to compensation data, correcting electromagnetic ripple propagation delay in body of wall, removing front wall affects interior wall.Construct a new matrix, make matrix also have feature block, successional, by matrix, scene is connected with projection, point is linked to be block, presents walls shape.The method operand is little such that it is able to be applied to real-time testing system.

Description

A kind of building layout formation method and system

Technical field

The present invention relates to through-wall radar imaging technical field, be specifically related to a kind of building layout formation method and system.

Background technology

Existing thunder imaging through walls not only considers to be hidden in the point target after body of wall based on human body, the most also to consider building The Extended target such as the body of wall within thing.Study the body of wall layout at interior of building most important in through-wall radar imaging, it Foundation is provided for the war against terror, hostage's rescue and urban SOS.

It is that intensive layout bay collection data are right outside building masonry wall that existing structure is laid out in image space fado The total evidence of post processing, to body of wall imaging after frequency domain direct imaging method processes.But, there is following asking in existing formation method Topic:

1, different due to thickness and the dielectric constant of front wall, electromagnetic wave propagation time delay can be produced impact, thus change Electromagnetic wave propagation path and speed, introduce target scattering echo delay error, and then occur that position skew, image diffusion etc. are existing As so that image quality is the best.

2, interior wall distance aerial position is different, then scattering point value differs, if imaging algorithm does not consider this of echo-signal The impact on scene imaging of one feature, will cause target clutter ratio little, and imaging effect also can be made the best.

3, for obtaining preferable imaging results, need antenna spacing little, and process total evidence, thus result in rear terminal number Big according to processing pressure.

In the Through-Wall Radar System of existing Step Frequency system, because transmitting terminal works asynchronously with receiving terminal, so existing Direct wave problem.Direct wave is signal the strongest between dual-mode antenna, not only the dynamic range of system for restricting, and when this signal Inversefouriertransform, to after time domain, has a rise time the shortest and produces the response of a sinc function shape, having Distance side lobe, for quiescent imaging, this response is equivalent to noise, can cover apart from upper neighbouring Small object.

Summary of the invention

The technical problem to be solved is, array element data sparse sampling situation big in the array element distance of real array Under, there is image quality difference in existing structure layout formation method and Step Frequency radar system exists stronger going directly Ripple, it is provided that a kind of building layout formation method and system.

For solving the problems referred to above, the present invention is achieved by the following technical solutions:

A kind of building layout formation method, comprises the steps:

Step 1, use the stepped frequency signal of m frequency as the transmitting signal of radar system；

Step 2, by n bay composition linear array, all antennas are h, every 2 antenna phases with the distance of exterior wall Spacing distance is L mutually；Antenna is set to bistatic mode of operation；

Step 3, aerial array integral translation is measured；Often one position measurement of translation is once, each translation position With an antenna for launching antenna, then remaining n-1 antenna is as reception antenna, the echo data obtained by each translation position Matrix, as echo data matrix v=[v obtained behind k position of translation₁,v₂,...,v_N]；Wherein N=(n-1) × k；

Step 4, the echo data matrix obtaining each translation position carry out inversefouriertransform, obtain one-dimensional echo letter Number, and one-dimensional echo-signal rising edge starting position is measured as inside the body of wall of each translation position to trailing edge end position Propagation delay τ_w；

The body of wall internal communication delay, τ that step 5, basis obtain_w, antenna and body of wall geometry of position relation, combine solve N number of Equation group estimates thickness d and the DIELECTRIC CONSTANT ε of body of wall；

Step 6, randomly choose in N group echo data S group antenna combination data and to its construct sparse dictionary ψ, its Middle S < N；

Step 7, structure gaussian random calculation matrix Φ=diag{ β₁,β₂,...β_S}；Wherein β_uDimension be H row, m arrange, And H < m；β_uIn element independently to obey average be 0, variance isGauss distribution；And according to z_u=β_uv_uObtain measurement data VectorWherein u=1,2 ..., S；v_uIt is the u column vector of echo data matrix v；

Step 8, structure G_x× B row, G_x×G_zThe block matrix A of row；(the q-1) × l of the q row of this block matrix A_x+ 1～q × l_xIndividual element is 1, and other element is 0；Above-mentioned G_xFor orientation to grid division number, G_zFor distance to grid division number, l_xFor definition Block length, B be orientation to block number,

Step 9, pseudoinverse is asked to obtain A-block matrix A¹, and it is constructed the dictionary matrix σ made new advances, i.e. σ with sparse dictionary ψ =ψ × A^-1；

Step 10, gaussian random calculation matrix Φ and new dictionary matrix σ are constructed sensing matrix θ, i.e. θ=Φ × σ；

Step 11, under constraints z=θ r, solve r；And try to achieve g according to Ag=r；It is derived from body of wall imaging knot Really.

In step 1, used stepped frequency signal is a string single carrier frequency pulse, and interpulse frequency is with step frequency △ f All hook incremental signal.

In step 5, the N number of equation group solved is:

$\left\{\begin{array}{c}2\frac{\sqrt{x_e^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_e^2+d^2}}{c}={\mathrm{\&tau;}}_{we}\\ \frac{x_e\sqrt{y_e^2+d^2}}{y_e\sqrt{x_e^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.$

In formula, x_eIt is the e electromagnetic wave incident point and the corresponding horizontal range launching antenna null position, y_eIt is the e electricity Magnetic wave incidence point and the e vertical horizontal range to body of wall front surface of refraction point, antenna spacing L=x_e+y_e, τ_weFor in body of wall Portion's propagation delay, e=1,2 ..., N, N=(n-1) × k, h is sky line-spacing front wall distance, and c is propagation velocity of electromagnetic wave, and d is wall The thickness of body, ε is the dielectric constant of body of wall.

In step 6, the detailed process of structure sparse dictionary ψ is:

Step 6.1, to imaging region discretization, imaging region is divided into G_x×G_zIndividual pixel grid；

Step 6.2, constructing ellipse according to the wave character of the one-dimensional echo-signal in step 4, N group dual-mode antenna combines structure Produce N number of ellipse, and the elliptic domain of structure is mapped in the pixel grid of division；

Thickness of wall body d and DIELECTRIC CONSTANT ε compensation calculation focusing time delay that step 5 solves is quoted in step 6.3, elliptic domain, Elliptic domain is outer calculates focusing time delay according to free space, thus builds complete sparse dictionary matrix ψ.

In step 6.3,

Focusing time delay τ in elliptic domain_inFor:

${\mathrm{\&tau;}}_{in}=\frac{2\&times;\left(\sqrt{{({\mathrm{\&mu;}}_{tn}-X_p)}^2+{({\mathrm{\&mu;}}_{rn}-Y_p)}^2}\right)}{c}+\frac{2\&times;(d\&times;(\sqrt{\mathrm{\&epsiv;}}-1))}{c}$

Focusing time delay τ outside elliptic domain_outFor:

${\mathrm{\&tau;}}_{out}=\frac{2\&times;\left(\sqrt{{({\mathrm{\&mu;}}_{tn}-X_p)}^2+{({\mathrm{\&mu;}}_{rn}-Y_p)}^2}\right)}{c}$

In formula, μ_tnRepresent and launch antenna coordinate position；μ_rnRepresent reception antenna coordinate position, (X_p, Y_p) represent pth picture Element mesh point.

In step 6.3, imaging region is divided into free space portion (1～G before three parts, i.e. body of wall_f), wall part (G_f+ 1～G_b) and body of wall after free space portion (G_b+ 1～G_xG_z), thus building complete dictionary ψ is:

$\begin{array}{c}{\&lsqb;{\mathrm{\&psi;}}_u\&rsqb;}_i=\left\{{\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i\right\}\\ =\left\{\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(Gf\right)}\end{array}\&rsqb;,\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\&rsqb;,\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{C}_z\right)}\end{array}\&rsqb;\right\}\end{array}$

In formula,For free space dictionary before body of wall,For body of wall dictionary,For the most empty after body of wall Between dictionary, ω_iFor step frequency, τ_inFor the focusing time delay in elliptic domain, τ_outFocusing time delay outside for elliptic domain, i=1,2, 3...m, u=1,2,3...S.

In step 5, dictionary ψ only considers that, just to antenna case, when grid position is just to antenna, ξ is 1, is otherwise 0, this Time,

Before body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& ...& \mathrm{\&xi;}(G_f,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_f\right)}\end{array}\&rsqb;$

Body of wall dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(G_f+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& ...& \mathrm{\&xi;}(G_b,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\&rsqb;$

After body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(G_b+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& ...& \mathrm{\&xi;}(G_x{G}_z,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{G}_z\right)}\end{array}\&rsqb;$

In formula,For free space dictionary before body of wall,For body of wall dictionary,For the most empty after body of wall Between dictionary, ω_iFor step frequency, τ_inFor the focusing time delay in elliptic domain, τ_outFocusing time delay outside for elliptic domain, i=1,2, 3...m, u=1,2,3...S.

Based on a kind of building layout imaging system designed by said method, including microprocessor, vector network analysis Instrument, microwave radio switching network, linear antenna arrays and host computer；

Linear antenna arrays: be made up of multiple electromagnetic horns, selects transmitting-receiving sky by the control of microwave radio switching network Line combines, it is achieved the mode of operation of electromagnetic horn multicast；

Microwave radio switching network: be located between vector network analyzer and linear antenna arrays, selects including transmitting chain Select circuit, receive walking circuit and antenna condition selection circuit；Wherein the transmitting terminal of vector network analyzer connects transmitting Walking circuit；The receiving terminal of vector network analyzer connects reception walking circuit；Antenna condition selection circuit connects Between transmitting chain selection circuit and linear antenna arrays, and receive between walking circuit and linear antenna arrays, Realize the conversion of antenna reiving/transmitting state in linear antenna arrays；Under control of the microprocessor, it is achieved transmitter and receiver divides Time mode of operation,

Vector network analyzer: realize the transmitting of signal and the collection of data；

Microprocessor: realize the control of the system work of vector network analyzer and microwave radio switching network；

Host computer: realize vector network analyzer and the control of microprocessor, simultaneously to vector network analyzer collection To data preserve and process, to realize the imaging of building layout.

The present invention, first with the geometry site of body of wall Yu antenna, estimates body of wall parameter, calculates electromagnetic wave and penetrates wall Body time delay, is used as follow-up imaging compensating by the time delay truly penetrating body of wall.Body of wall quantity in whole building interior space Few, have openness, utilize body of wall to have bulk, successional feature simultaneously.Use distributed compression sensing method, to target Distance domain divide subspace, by calculated true propagation time delay in subspace to compensation data, correcting electromagnetic ripple is at wall Propagation delay in body, removing front wall affects interior wall.Construct a new matrix A, make matrix A also have bulk, successional Feature, is connected scene with projection by A matrix, point is linked to be block, presents walls shape.The method operand is little, from And real-time testing system can be applied to.

Compared with prior art, the present invention has a characteristic that

1, in algorithm part, first according to position and the body of wall geometry site of dual-mode antenna, measurement time delay is utilized to estimate Meter body of wall parameter；Secondly then utilize one-dimensional echo character to build elliptic domain, and combine elliptic domain structure dictionary, in elliptic domain Introducing body of wall parameter and calculate focusing time delay, elliptic domain is outer calculates focusing time delay according to free space；Special again according to body of wall bulk Property build a matrix comprising body of wall position in advance, this matrix also has block characteristic；Then the original word of this matrix update is utilized Allusion quotation matrix, finally uses recovery algorithms imaging.

2, design part in system, utilize transmitting chain and the reception link asynchronous working mould of microwave radio switching network Formula so that emitting portion and receiving portion time-sharing work, reaches to suppress antenna direct wave purpose.

Accompanying drawing explanation

Fig. 1 is the schematic diagram that geometric model estimates body of wall parameter.

Fig. 2 is the schematic diagram that elliptic domain is set up.

Fig. 3 is compressed sensing framework schematic diagram.

Fig. 4 is the schematic diagram that scene grid is set up.

Fig. 5 (a) is A Matrix Construction Method schematic diagram, and Fig. 5 (b) is A matrix structure schematic diagram.

Fig. 6 (a) is traditional method imaging results, and Fig. 6 (b) is context of methods imaging results.

Fig. 7 is the theory diagram of a kind of building layout imaging system.

Fig. 8 is the schematic diagram of microwave radio switching network.

Detailed description of the invention

A kind of building layout formation method, comprises the steps:

Step 1: using stepped frequency signal as the transmitting signal of radar system, its basic thought is: launch a string successively (m) pulse width is T_pSingle carrier frequency pulse, interpulse frequency all hooks incremental with △ f, and it launches signal expression For:Wherein, f₀For initial frequency, △ f is step frequency,For initial phase, stepping Count i=1,2 ..., m.

Step 2: by n bay composition linear array, antenna distance exterior wall distance is h, and each bay is mutual Spacing distance is L, and antenna is set to bistatic mode of operation.One antenna is for launching antenna, then remaining n-1 antenna As reception antenna, the corresponding echo data gathering the combination of n-1 group dual-mode antenna.Because use stepped frequency signal, corresponding echo There is m frequency, so the data form of storage is for having m row, the matrix of n-1 row.

Step 3: because the linear array synthetic aperture of n bay composition is limited, it is impossible to completely in detection building Portion's body of wall distribution situation, so needing to measure aerial array integral translation.Often one position measurement of translation is once, when flat After moving k position, be allowed to outside the building enclose and can detect completely, have k × (n-1) group dual-mode antenna combination, gather accordingly k × (n-1) individual echo data.Convenient for representing, be designated as N, i.e. N=(n-1) × k, now add up to N group echo data, then echo-signal For v=[v₁,v₂,…,v_N], echo data matrix existing m row N arranges, and is expressed as C^m×N。

Step 4: echo-signal is carried out inversefouriertransform, obtains the one-dimensional echo-signal of target, owing to electromagnetic wave runs into Target can produce peak value, therefore also can produce peak value, according to wave character, by echo rising edge start bit when running into body of wall reflection Put trailing edge end position and be measured as body of wall internal communication delay, τ_w。

Step 5: electromagnetic wave can change electromagnetic wave propagation path and speed at body of wall internal communication, if not considering front wall pair The impact of electromagnetic wave will introduce target scattering echo delay error, and then the phenomenons such as position skew, image diffusion occurs so that Image quality is the best.Different body of wall dielectric constants or different thickness can be different on the impact of electromagnetic wave propagation time delay.And In actual application, body of wall unknown parameters, so this step according to step 4 measure that the electromagnetic wave obtained penetrates inside body of wall time Prolong τ_w, also antenna and body of wall geometry of position relation, combine and solve multiple equation and estimate thickness of wall body d and DIELECTRIC CONSTANT ε.Make Compensate in the dictionary of subsequent step constructs by body of wall parameter, thus correct body of wall position and suppression image diffusing phenomenon.

Step 6: under compressed sensing framework, uses low volume data can revert to picture, at N group echo data v=[v₁, v₂,...,v_NRandomly choose the data of S group (S < N) antenna combination in] and it is constructed sparse dictionary, i.e. to signal v=[v₁, v₂,…,v_S] structure dictionary, only revert to picture by S group data.The echo that the u group dual-mode antenna received can be combined Signal v_uIt is expressed as dictionary matrix ψ_uLinear combination with multiple point target reflection coefficient vector r: v_u=ψ_uR, u=1,2, 3...S.First to imaging region discretization, dividing pixel grid, orientation is to dividing G into_xIndividual pixel grid, distance is to being divided into G_z Individual pixel grid, a total of P=G_x×G_zIndividual pixel grid.Then structure dictionary matrix ψ_u, believe according to the one-dimensional echo of step 4 Number wave character, be oval semi-minor axis a with wide 1/2 distance of ripple, with two center of antenna distances L for ellipse semi-major axis b, structure Make ellipse.Echo data relative configurations owing to there being N group dual-mode antenna to combine goes out N number of ellipse, is mapped to by the elliptic domain of structure In the pixel grid divided.The body of wall parameter compensation calculation focusing time delay that step 5 solves is quoted in elliptic domain；Press outside elliptic domain Focusing time delay is calculated according to free space.

Step 7, structure gaussian random calculation matrix Φ=diag{ β₁,β₂,...β_S}；Wherein β_uDimension be H row, m arrange, And H < m；β_uIn element independently to obey average be 0, variance isGauss distribution；And according to z_u=β_uv_uObtain measurement data VectorWherein u=1,2 ..., S；v_uIt is the u column vector of echo data matrix v.

Step 8, structure G_x× B row, G_x×G_zThe block matrix A of row；(the q-1) × l of the q row of this block matrix A_x+ 1～q × l_xIndividual element is 1, and other element is 0；Above-mentioned G_xFor orientation to grid division number, G_zFor distance to grid division number, l_xFor definition Block length, B be orientation to block number,

Step 9, pseudoinverse is asked to obtain A block matrix A^-1, and it is constructed the dictionary matrix σ made new advances, i.e. σ with sparse dictionary ψ =ψ × A^-1。

Step 10, gaussian random calculation matrix Φ and new dictionary matrix σ are constructed sensing matrix θ, i.e. θ=Φ × σ.

Step 11, under constraints z=θ r, solve r；And try to achieve g according to Ag=r；It is derived from body of wall imaging knot Really.

Illustrate with a body of wall:

By n bay composition linear array, bay be spaced apart L, sky line-spacing front wall distance h.Launch sky for 1 Line, n-1 reception antenna, should there is the echo data that n-1 group dual-mode antenna combines mutually, by mobile k position, gather (n-altogether 1) × k echo data, is designated as N, wherein N=(n-1) × k.By antenna and body of wall position, and electromagnetic wave propagation path is several Body of wall estimated by what model, utilizes the mode of operation of multicast, mobile k position, gathers N group data.According to electromagnetic wave propagation Time delay, by solving multiple equation group, optimization, estimates body of wall parameter, including thickness of wall body d and dielectric constant ε.As it is shown in figure 1, electromagnetic wave propagation meets snell law, r₁Represent that electromagnetic wave enters into the incidence of body of wall through free space Point position, R₁Represent refraction point position, body of wall rear surface, r₂Represent the reflection point position from body of wall to free space, x₁Represent the Launch antenna null position to r for one₁Horizontal range length, y₁Represent r₁To R₁The vertical horizontal range to body of wall front surface Length, d represent thickness of wall body, h represent antenna to front wall distance, L represents antenna spacing distance.

$2\frac{\sqrt{x_1^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_1^2+d^2}}{c}={\mathrm{\&tau;}}_{w_1}$

WhereinWherein L=x₁+y₁, with above-mentioned two equation equationof structure group, it is known that amount has measurement Delay, τ_w1, antenna spacing L, sky line-spacing front wall distance h, propagation velocity of electromagnetic wave c, equation group has 2 equations, but has 3 Unknown quantity, including electromagnetic wave incident point and corresponding horizontal range x launching antenna null position₁, thickness of wall body d, body of wall dielectric Constant ε, therefore cannot solve.But be introduced into the data of the 2nd reception antenna, then equation group is:

$\left\{\begin{array}{c}2\frac{\sqrt{x_1^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_1^2+d^2}}{c}={\mathrm{\&tau;}}_{w1}\\ \frac{x_1\sqrt{y_1^2+d^2}}{y_1\sqrt{x_1^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.$

$\left\{\begin{array}{c}2\frac{\sqrt{x_1^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_1^2+d^2}}{c}={\mathrm{\&tau;}}_{w2}\\ \frac{x_2\sqrt{y_2^2+d^2}}{y_2\sqrt{x_2^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.$

Above-mentioned equation group comprises 4 equations, and unknown quantity includes the electromagnetic wave incident point position x of first group of antenna₁, second group Antenna electric magnetic wave incidence point and corresponding horizontal range x launching antenna null position₂, thickness of wall body d and DIELECTRIC CONSTANT ε, for Seek optimal solution, the data of introducing N group dual-mode antenna combination:

$\begin{array}{c}\left\{\begin{array}{c}2\frac{\sqrt{x_1^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_1^2+d^2}}{c}={\mathrm{\&tau;}}_{w1}\\ \frac{x_1\sqrt{y_1^2+d^2}}{y_1\sqrt{x_1^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.\\ \left\{\begin{array}{c}2\frac{\sqrt{x_2^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_2^2+d^2}}{c}={\mathrm{\&tau;}}_{w2}\\ \frac{x_2\sqrt{y_2^2+d^2}}{y_2\sqrt{x_2^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.\\ \mathrm{......}\\ \left\{\begin{array}{c}2\frac{\sqrt{x_N^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_N^2+d^2}}{c}={\mathrm{\&tau;}}_{wN}\\ \frac{x_N\sqrt{y_N^2+d^2}}{y_N\sqrt{x_N^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.\end{array}$

Above-mentioned total 2N equation, N+2 unknown quantity, unknown quantity includes N number of electromagnetic wave incident point and corresponding transmitting antenna Horizontal range x of dead-center position₁,x₂......x_N, thickness of wall body d and DIELECTRIC CONSTANT ε, the quantity of equation is much larger than unknown quantity Quantity, so equation can solve, eventually through measuring time delay and estimating the lowest mean square of the calculated estimation time delay of ε and d by mistake Difference obtains optimum d and ε.

Elliptic domain is set up with above-mentioned echo waveform feature, as in figure 2 it is shown, with a body of wall citing, Fu anti-to echo-signal In leaf transformation, peak value can be produced owing to electromagnetic wave runs into target, also can produce peak value, at peak value generation i.e. when running into body of wall reflection For body of wall, do not produce at peak value i.e. driftlessness, for free space, utilize wave character to construct elliptic domain, with 1/2 distance that ripple is wide For oval semi-minor axis a, with two center of antenna distances L for oval semi-major axis b, structure ellipse, the combination of N group dual-mode antenna has N number of The oval elliptic domain formed.

Under the compressed sensing framework shown in Fig. 3, use low volume data can revert to picture, in N group echo data with Machine selects the data of S group (S < N) antenna combination and it is constructed sparse dictionary.In imaging algorithm processes, need imaging area Territory divides pixel grid, and orientation is to being divided into G_xIndividual pixel grid, distance is to being divided into G_zIndividual pixel grid, a total of P=of scene G_x×G_zIndividual pixel grid.Elliptic domain is mapped to pixel grid, as shown in Figure 4.In elliptic domain, dictionary introduces the calculating of body of wall parameter Focusing time delay, elliptic domain is outer calculates focusing time delay according to free space.After above-mentioned elliptic domain is mapped to grid, image scene i.e. quilt Being divided into free space portion before three parts, i.e. body of wall, grid scope includes (1～G_f)；Wall part, grid scope includes (G_f + 1～G_b) and body of wall after free space a part, grid scope includes (G_b+ 1～G_xG_z).Three are divided according to above-mentioned image scene Part, dictionary is according to scene grid distributed structure, to S its sparse dictionary of group data configuration, wherein

Before body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& ...& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_f\right)}\end{array}\&rsqb;,i=1,2,\mathrm{3...}m,u=1,2,\mathrm{3...}S$

In the case of GaiFor elliptic domain exterior focusing time delay, calculate N the most one by one The position of group transmitting-receiving combined antenna is to (1～G_f) electromagnetic wave propagation time of mesh point, it is focusing time delay, μ_tnRepresent and launch sky Line coordinates position, μ_rnRepresent reception antenna coordinate position, free space portion before now mesh point is elliptic domain outer grid, i.e. body of wall Subnetting lattice, wherein ω_i=f₀₊I × △ f, i=1,2,3...m.

Body of wall dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& ...& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\&rsqb;,i=1,2,\mathrm{3...}m,u=1,2,\mathrm{3...}S$

NowFormula Section 2 is cited as compensating , introducing the impact of thickness of wall body d and DIELECTRIC CONSTANT ε compensation body of wall, c represents that the light velocity is the brightest 3 × 10⁸m/s；Count the most one by one Calculate transmitting antenna and then arrive the electromagnetic wave propagation time of reception antenna, grid scope (G to each mesh point in elliptic domain_f+ 1～ G_b), wherein ω_i=f₀+ i × △ f, i=1,2,3...m, n=1,2,3...N.

After body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{G}_z\right)}\end{array}\&rsqb;,i=1,2,3,\mathrm{...},m,u,=1,2,\mathrm{3...}S$

NowFor elliptic domain exterior focusing time delay, calculate transmitting-receiving the most one by one The position of antenna combination is to (G_b+ 1～G_xG_z) electromagnetic wave propagation time of mesh point, now mesh point is the outer grid of elliptic domain, I.e. free space portion grid after body of wall.

Final merge distribution dictionary, build complete sparse dictionary:

${\&lsqb;{\mathrm{\&psi;}}_u\&rsqb;}_i=\left\{{\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i\right\},i=1,2,\mathrm{...},m,u=1,2,\mathrm{3...}S$

Only consider that when grid position is just to antenna, following formula ξ is 1 just to antenna case with the dictionary of above-mentioned structure；Otherwise It is 0.

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& ...& \mathrm{\&xi;}(G_f,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_f\right)}\end{array}\&rsqb;,i=1,2,\mathrm{3...}m,u=1,2,\mathrm{3...}S$

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(G_f+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& ...& \mathrm{\&xi;}(G_b,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\&rsqb;,i=1,2,3,\mathrm{...}m,u=1,2,\mathrm{3...}S$

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i=\&lsqb;\begin{array}{ccc}\mathrm{\&xi;}(G_b+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& \mathrm{...}& \mathrm{\&xi;}(G_x{G}_z,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{G}_z\right)}\end{array}\&rsqb;,i=1,2,3,\mathrm{...},m,u,=1,2,\mathrm{3...}S$

With above-mentioned pixel grid in orientation to being divided into block of pixels, block length is l_x, block number isBuild one newly Matrix A, its line number is determined by the block of pixels number that divides, and columns is determined by the pixel grid number divided.

A matrix becomes block feature so that it is pre-estimation body of wall position, every a line of A all may comprise in imaging one long Degree is l_xOne side body of wall.By A matrix, imaging is connected with scene.Original compressed sensing framework z=Φ ψ r, present z =Φ ψ Ag, vector g are target scattering coefficient.Fig. 5 (a) is A Matrix Construction Method schematic diagram, and Fig. 5 (b) is that A matrix structure shows It is intended to.

With above-mentioned A Matrix Calculating pseudo inverse matrix A-¹, build complete dictionary matrixSo Rear A^-1The dictionary matrix σ that structure is new, σ=ψ × A is contacted with dictionary matrix ψ^-1, construct gaussian random calculation matrix Φ, Gao Sisui Machine calculation matrix Φ constructs sensing matrix θ, θ=Φ × σ with new dictionary matrix σ.Final compressed sensing equation is updated to: z= ΦψA^-1R=Φ σ r=θ r.Under constraints z=θ r, solve r；And try to achieve g according to Ag=r；It is derived from body of wall imaging Result.

Fig. 6 (a) is to use traditional method imaging results, and Fig. 6 (b) is to use context of methods imaging results.One has 21 groups The echo data of dual-mode antenna combination, each antenna spacing 30cm, meet array element openness.Algorithm is only with 10 groups of transmitting-receivings herein Antenna combination data reconstruction obtains high-definition picture.Relative to traditional algorithm, discrete point is linked to be block by the present invention, and makes Obtaining the scattering point value in same identical, meeting body of wall is bulk, successional feature, and utilizes compressed sensing framework, real Showed bay sparse in the case of high-resolution imaging.

The present invention, first according to position and the body of wall geometry site of the combination of N group dual-mode antenna, utilizes measurement time delay to estimate Meter body of wall parameter；Secondly then utilize one-dimensional echo character to build elliptic domain, and combine elliptic domain structure dictionary, in elliptic domain Introducing body of wall parameter and calculate focusing time delay, elliptic domain is outer calculates focusing time delay according to free space；Special again according to body of wall bulk Property build a matrix comprising body of wall position in advance, this matrix also has block characteristic；Then the original word of this matrix update is utilized Allusion quotation matrix, finally uses recovery algorithms imaging.

Based on a kind of building layout imaging system designed by said method, as it is shown in fig. 7, comprises vector network analysis Instrument, microwave radio switching network, microprocessor and ultra broadband electromagnetic horn.

Vector network analyzer: be mainly used in the performance test of microwave device, is widely used in antenna measurement, circuit survey The fields such as examination, component testing and measurement verification.Vector network analyzer can be used to measure passive and active electric network S parameter, It is double (or four) passage microwave receiver, is designed to the width of transmitted wave and the echo that can be used to process automatic network Value and phase place.Utilize the characteristic of vector network analyzer, it would be desirable to the target scene of detection is considered as Measurement Network, be incorporated into through walls In radar imagery, PC realizes control to vector network analyzer by LabVIEW host computer, using this instrument as emission source and Receiver, it is achieved transmitting and the collection of data to signal, is saved in host computer, and LabVIEW uses dual-thread to operate, one Thread gathers data, and the data being stored on PC, based on embedded mathscript module, are processed by another thread, logical Cross algorithm and realize building layout imaging, reach the method for work of online treatment data imaging, and tie in interface display imaging Really.The vector network analyzer that Based PC host computer LabVIEW controls, sends/accepts radiofrequency signal, measures body of wall and hides The amplitude of target and phase place, at PC host computer collection signal processing, show at LabVIEW and be ultimately imaged result.

Microwave radio switching network: use microwave radio switching network between vector network analyzer and antenna, not only will Transmitting terminal and the extension of receiving terminal passage, it is achieved the synthetic aperture required for the linear array of multiple bays composition, make simultaneously System method of work is similar to time domain system.Transmitting terminal at vector network analyzer connects transmitting chain selection circuit；Receiving End connects reception walking circuit；Transmitting chain selection circuit/be connected antenna-like between reception walking circuit and antenna State change-over circuit, it is possible to control the conversion of antenna reiving/transmitting state.Signal source port, uses microwave to penetrate between receiving port and antenna Frequently switching network, utilizes switching network asynchronous work mode so that transmitter and receiver time-sharing work, reaches to suppress coupled wave Purpose.Above circuit module uses microprocessor (MCU) to control, and link channel selects to control end write by MCU to it and compiles Code signal realizes, and Systematic selection realizes the mode of operation of multicast, makes system simultaneously launching signal when, receives End is closed；The when that receiving terminal receiving signal, transmitting terminal is closed.

Microwave radio switching network is illustrated with 6 electromagnetic horns of extension.Described microwave radio switching network such as Fig. 8 institute Showing, specifically by switch chip, power amplifier, low-noise amplifier, capacitance forms.It is divided into transmitting chain, receives link, And antenna reiving/transmitting state modular converter.Transmitting chain is identical with receiving link structure, uses SP8T chip, antenna reiving/transmitting state Conversion is realized by SP3T chip.The port1 of vector network analyzer connects transmitting chain, connects through 100pF capacitance and sends out Penetrating the RFC end of link SP8T1, RF1～the RF6 end of SP8T1 connects the input of power amplifier by 100pF capacitance, The outfan of power amplifier connects the RF1 end of SP3T1.RF7 and the RF8 end of transmitting chain SP8T1 is through 50 Ω load resistances Connect GND end.The port2 of vector network analyzer connects reception link, connects through 100pF capacitance and receives link The RFC end of SP8T2, RF1～the RF6 end receiving link SP8T connects the output of low-noise amplifier by 100pF capacitance End, the input of low-noise amplifier connects the RF2 end of SP3T2.The RFC end of SP3T2 connects loudspeaker through the capacitance of 100pF Antenna terminal, RF3 end connects GND end through 50 Ω load resistances.Transmitting chain SP8T1, receives continuous SP8T2, and antenna is received That sends out State Transferring SP3T controls the IO that end connects MCU, is controlled the channel selecting of switch chip by low and high level.Transmitting chain Power amplifier be that the power making system launch exceedes for compensating the duty-cycle loss caused under transmitting chain closed mode Do not use the possible maximum of switching network.The low-noise amplifier receiving link act as the noise coefficient of improvement system also Improve system gain.

Described microwave radio switching network suppression coupled wave method is with antenna 1 for launching antenna, and antenna 2 is lifted for reception antenna Example explanation.The ultra-broadband signal persistent period is γ.First, the RF1 conducting of SP8T1, the RF1 conducting of SP3T1, the RF3 of SP3T2 leads Logical, it is biased towards carrying ground, the RF7 conducting of SP8T2, it is biased towards carrying ground；Now, transmitting chain turns on, and transmitter works； Receiving link and be biased towards carrying ground, so receiving link is off-state, receiver does not works.Start to export ultra-wide to take a message Number, γ₁(γ after time₁=1.1 × γ), the RF7 conducting of SP8T1, it is biased towards carrying ground, the RF3 conducting of SP3T1, it is biased towards Carrying ground, the RF2 conducting of SP8T2, the RF2 conducting of SP3T2, receive link conducting, operation of receiver receives echo-signal.Logical Cross the Asynchronous operation mode of microwave radio switching network, make transmitter and receiver time-sharing work so that the method for work of system It is similar to time domain system, reaches to suppress the purpose of antenna coupled wave.This switching network simple in construction, it is easy to accomplish.

Linear antenna arrays: be made up of multiple antennas, by the transmitting chain selection circuit of microwave radio switching network, receives Walking circuit, antenna condition selection circuit forms.Dual-mode antenna is selected to combine by above circuit module, it is achieved array element one Send out the mode of operation received more.

6 antenna composition linear arrays are illustrated and indicate sequence number, select sequence number by described electromagnetic horn composition linear array 1 antenna for launching antenna, then selects then sequence number 2 by the reception link of microwave radio switching network successively, and 3,4,5,6 antennas are Reception antenna, is expressed as [1,2], [1,3], [Isosorbide-5-Nitrae], [1,5], [1,6], then be 5 kinds of dual-mode antenna compound modes.

Realize passage by microwave radio switching network extend and realize reiving/transmitting state conversion, it is achieved the MIMO of MIMO Antenna configurations.

The work process of a kind of building layout imaging system is as follows:

Step 1: configuration Instrument working parameter.

Step 2: launch the transmitting chain conducting of antenna；The reception link of reception antenna is biased towards carrying ground.

Step 3: launch antenna output radiofrequency signal.

Step 4: radiofrequency signal output is complete.The transmitting chain launching antenna is biased towards carrying ground；Connecing of reception antenna Receipts link turns on.

Step 5: host computer LabVIEW dual-thread works, and a thread gathers the data of next antenna combination；Another line The echo data that receives is used a kind of building layout formation method to process in real time by journey, then switching position repeat step 2, 3、4。

Step 6: repeat step 5.

Native system builds building layout sparse snapshot imaging Auto-Test System with vector network analyzer for core, profit By the characteristic of vector network analyzer, it would be desirable to the target scene of detection is considered as Measurement Network, is incorporated in through-wall radar imaging. With vector network analyzer as core, measure body of wall and the amplitude of vanishing target and phase place, write certainly by host computer LabVIEW Dynamic control program, it is achieved the mode of operation of data acquisition and procession imaging integration, LabVIEW program dual-thread operates, one Thread gathers data, and another thread is processed into picture to the data gathered, and meets requirement of real-time, it is achieved building layout is online The function of imaging.Microwave radio switching network is used to realize the extension to port number, it is achieved joining of MIMO (MIMO) array Put, be simultaneous for frequency domain radar and there is antenna coupled wave problem, utilize microwave radio switching channels to switch, reach transmitting chain and Receiving the purpose of link asynchronous working, therefore, the application of this switching network is equivalent to pulse modulation method so that system work side Method is similar to time domain radar, and transmitter works when, receiver is forbidden receiving, and the when of operation of receiver, transmitter is forbidden Launching, in switching network, the conversion time launched/receive is faster than the frequency step time, causes pulse modulation frequency to export, The application of this switching network can effectively suppress antenna coupled wave.

Claims (8)

1. a building layout formation method, it is characterised in that: comprise the steps:

Step 1, use the stepped frequency signal of m frequency as the transmitting signal of radar system；

Step 2, by n bay composition linear array, all antennas are h with the distance of exterior wall, and every 2 antennas are mutual Space from for L；Antenna is set to bistatic mode of operation；

Step 3, aerial array integral translation is measured；Often one position measurement of translation is once, and each translation position is with one Individual antenna is for launching antenna, then remaining n-1 antenna is as reception antenna, the echo data square obtained by each translation position Battle array, as echo data matrix v=[v obtained behind k position of translation₁,v₂,...,v_N]；Wherein N=(n-1) × k；

Step 4, the echo data matrix obtaining each translation position carry out inversefouriertransform, obtain one-dimensional echo-signal, And the internal biography of body of wall that one-dimensional echo-signal rising edge starting position to trailing edge end position is measured as each translation position Broadcast delay, τ_w；

The body of wall internal communication delay, τ that step 5, basis obtain_w, antenna and body of wall geometry of position relation, combine and solve N number of equation Group estimates thickness d and the DIELECTRIC CONSTANT ε of body of wall；

Step 6, randomly choose in N group echo data S group antenna combination data and to its construct sparse dictionary ψ, wherein S < N；

Step 7, structure gaussian random calculation matrix Φ=diag{ β₁,β₂,...β_S}；Wherein β_uDimension be H row, m arrange, and H < m；β_uIn element independently to obey average be 0, variance isGauss distribution；And according to z_u=β_uv_uObtain measurement data vectorWherein u=1,2 ..., S；v_uIt is the u column vector of echo data matrix v；

Step 8, structure G_x× B row, G_x×G_zThe block matrix A of row；(the q-1) × l of the q row of this block matrix A_x+ 1～q × l_xIndividual Element is 1, and other element is 0；Above-mentioned G_xFor orientation to grid division number, G_zFor distance to grid division number, l_xFor definition block length Degree, B be orientation to block number,

Step 9, pseudoinverse is asked to obtain A block matrix A^-1, and the dictionary matrix σ that it is made new advances with sparse dictionary ψ structure, i.e. σ=ψ × A^-1；

Step 10, gaussian random calculation matrix Φ and new dictionary matrix σ are constructed sensing matrix θ, i.e. θ=Φ × σ；

Step 11, under constraints z=θ r, solve r；And try to achieve g according to Ag=r；It is derived from body of wall imaging results.

A kind of building layout formation method the most according to claim 1, it is characterised in that: in step 1, used stepping Frequently signal is a string single carrier frequency pulse, and interpulse frequency is all to hook incremental signal with step frequency △ f.

A kind of building layout formation method the most according to claim 1, it is characterised in that: in step 5, the N solved Individual equation group is:

$\left\{\begin{array}{c}2\frac{\sqrt{x_e^2+h^2}}{c}+2\frac{\sqrt{\mathrm{\&epsiv;}}\sqrt{y_e^2+d^2}}{c}={\mathrm{\&tau;}}_{we}\\ \frac{x_e\sqrt{y_e^2+d^2}}{y_e\sqrt{x_e^2+h^2}}=\sqrt{\mathrm{\&epsiv;}}\end{array}\right.$

In formula, x_eIt is the e electromagnetic wave incident point and the corresponding horizontal range launching antenna null position, y_eIt is the e electromagnetic wave Incidence point and the e vertical horizontal range to body of wall front surface of refraction point, antenna spacing L=x_e+y_e, τ_wePass for body of wall inside Sowing time prolongs, e=1, and 2 ..., N, N=(n-1) × k, h is sky line-spacing front wall distance, and c is propagation velocity of electromagnetic wave, and d is body of wall Thickness, ε is the dielectric constant of body of wall.

A kind of building layout formation method the most according to claim 1, it is characterised in that: in step 6, construct sparse word The detailed process of allusion quotation ψ is:

Step 6.1, to imaging region discretization, imaging region is divided into G_x×G_zIndividual pixel grid；

Step 6.2, constructing ellipse according to the wave character of the one-dimensional echo-signal in step 4, N group dual-mode antenna composite construction goes out N number of ellipse, and the elliptic domain of structure is mapped in the pixel grid of division；

Thickness of wall body d and DIELECTRIC CONSTANT ε compensation calculation focusing time delay that step 5 solves is quoted in step 6.3, elliptic domain, oval Overseas according to free space calculating focusing time delay, thus build complete sparse dictionary matrix ψ.

A kind of building layout formation method the most according to claim 4, it is characterised in that: in step 6.3,

Focusing time delay τ in elliptic domain_inFor:

${\mathrm{\&tau;}}_{in}=\frac{2\&times;\left(\sqrt{{({\mathrm{\&mu;}}_{tn}-X_p)}^2+{({\mathrm{\&mu;}}_{rn}-Y_p)}^2}\right)}{c}+\frac{2\&times;(d\&times;(\sqrt{\mathrm{\&epsiv;}}-1))}{c}$

Focusing time delay τ outside elliptic domain_outFor:

${\mathrm{\&tau;}}_{out}=\frac{2\&times;\left(\sqrt{{({\mathrm{\&mu;}}_{tn}-X_p)}^2+{({\mathrm{\&mu;}}_{rn}-Y_p)}^2}\right)}{c}$

In formula, μ_tnRepresent and launch antenna coordinate position；μ_rnRepresent reception antenna coordinate position, (X_p, Y_p) represent pth pixel network Lattice point.

A kind of building layout formation method the most according to claim 4, it is characterised in that: in step 6.3, imaging region It is divided into free space portion (1～G before three parts, i.e. body of wall_f), wall part (G_f+ 1～G_b) and body of wall after free space Partly (G_b+ 1～G_xG_z), thus building complete dictionary ψ is:

$\begin{array}{c}{\&lsqb;{\mathrm{\&psi;}}_u\&rsqb;}_i=\{{\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i,{\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i\}\\ =\left\{\left[\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(Gf\right)}\end{array}\right],\left[\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\right],\left[\begin{array}{ccc}{e}^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& \mathrm{...}& e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{G}_z\right)}\end{array}\right]\right\}\end{array}$

In formula,For free space dictionary before body of wall,For body of wall dictionary,For free space word after body of wall Allusion quotation, ω_iFor step frequency, τ_inFor the focusing time delay in elliptic domain, τ_outFocusing time delay outside for elliptic domain, i=1,2,3...m, U=1,2,3...S.

A kind of building layout formation method the most according to claim 4, it is characterised in that: in step 5, dictionary ψ only examines Considering just to antenna case, when grid position is just to antenna, ξ is 1, is otherwise 0, now,

Before body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(1\right)}\&rsqb;}_i=\left[\begin{array}{ccc}\mathrm{\&xi;}(1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(1\right)}& \mathrm{...}& \mathrm{\&xi;}(G_f,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_f\right)}\end{array}\right]$

Body of wall dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(2\right)}\&rsqb;}_i=\left[\begin{array}{ccc}\mathrm{\&xi;}(G_f+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}(Gf+1)}& \mathrm{...}& \mathrm{\&xi;}(G_b,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{in}\left(G_b\right)}\end{array}\right]$

After body of wall, free space dictionary is:

${\&lsqb;{\mathrm{\&psi;}}_u^{\left(3\right)}\&rsqb;}_i=\left[\begin{array}{ccc}\mathrm{\&xi;}(G_b+1,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}(Gb+1)}& \mathrm{...}& \mathrm{\&xi;}(G_x{G}_z,u)e^{-{\mathrm{j\&omega;}}_i{\mathrm{\&tau;}}_{out}\left(G_x{G}_z\right)}\end{array}\right]$

In formula,For free space dictionary before body of wall,For body of wall dictionary,For free space word after body of wall Allusion quotation, ω_iFor step frequency, τ_inFor the focusing time delay in elliptic domain, τ_outFocusing time delay outside for elliptic domain, i=1,2,3...m, U=1,2,3...S.

8., based on a kind of building layout imaging system designed by method described in claim 1, it is characterized in that, including micro-process Device, vector network analyzer, microwave radio switching network, linear antenna arrays and host computer；

Linear antenna arrays: be made up of multiple electromagnetic horns, selects dual-mode antenna group by the control of microwave radio switching network Close, it is achieved the mode of operation of electromagnetic horn multicast；

Microwave radio switching network: be located between vector network analyzer and linear antenna arrays, selects electricity including transmitting chain Road, receives walking circuit and antenna condition selection circuit；Wherein the transmitting terminal of vector network analyzer connects transmitting chain Selection circuit；The receiving terminal of vector network analyzer connects reception walking circuit；Antenna condition selection circuit is connected to send out Penetrate between walking circuit and linear antenna arrays, and receive between walking circuit and linear antenna arrays, it is achieved The conversion of antenna reiving/transmitting state in linear antenna arrays；Under control of the microprocessor, it is achieved transmitter and receiver timesharing work Operation mode,

Vector network analyzer: realize the transmitting of signal and the collection of data；

Microprocessor: realize the control of the system work of vector network analyzer and microwave radio switching network；

Host computer: realize vector network analyzer and the control of microprocessor, vector network analyzer is collected simultaneously Data preserve and process, to realize the imaging of building layout.